November 25, 2014 
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >
November 25, 2014 [00:01:34] *** snakenerd has quit IRC[00:01:52] *** Codex_ has quit IRC[00:02:48] <Black0range> Are there some kind soul that could explain to me why this doesn't work as a if? http://pastebin.com/JdsRQYvq[00:04:35] <derhass> Black0range: putting all the type issues aside: a bitwise not is not what you want[00:05:07] <Black0range> :'([00:05:37] <foobaz> if you a boolean values, you want a logical not[00:05:38] *** Kane has quit IRC[00:05:52] <Black0range> no i want the bitwise[00:06:00] <foobaz> why?[00:06:01] <derhass> no, you don't[00:06:01] *** TyrfingMjolnir has joined ##OpenGL[00:06:17] <Black0range> well ok no i don't then :P[00:06:32] <derhass> your code expects ~1 to be 0[00:06:36] <derhass> but it isn't[00:06:42] *** ibqn has quit IRC[00:06:43] <derhass> it is 2^32-2[00:06:48] <derhass> for 32bit unsigned[00:07:30] <Black0range> oh so int() doesn't convert that?[00:07:35] *** snakener1 has quit IRC[00:07:36] <derhass> what?[00:07:51] <derhass> I don't get your point[00:07:54] <Black0range> shoudln't "oneIsSet" be either 0 or 1 ?[00:08:03] <derhass> yeah[00:08:03] <Jaus> the convertion doesn't do what you think it does.[00:08:09] <foobaz> yes, that’s why you don’t want bitwise not[00:08:31] <foobaz> logical operations are for 0 or 1 values[00:08:32] <derhass> Jaus: the conversion isn't the main issue here[00:09:01] <Black0range> oh yeah wait, i'm retarded[00:09:10] <Black0range> you're right i do not want bitwise[00:09:17] <Jaus> I know, I am just making sure he doesn't think casting is a magic idea that forces data to be in the format you want it to be.[00:09:44] <derhass> that whole construction appears a bit over-complicated[00:09:53] *** PasNox has quit IRC[00:10:13] <Black0range> well so i have a square that i want to give borders[00:10:17] <Black0range> straight borders[00:10:33] <Jaus> What other websites or tutorials are similar to open.gl?[00:11:26] <slime> http://www.arcsynthesis.org/gltut/ is pretty good[00:11:43] <Jaus> I want to relearn modern OpenGL because my foundation is a bit weak.[00:11:50] <Black0range> diff = current coordinate - ceneter of cube[00:12:11] *** Dudi has quit IRC[00:13:30] *** PasNox has joined ##OpenGL[00:13:39] <Black0range> oh i made it work[00:15:49] *** Codex_ has joined ##OpenGL[00:17:32] *** PasNox has quit IRC[00:17:58] *** PasNox has joined ##OpenGL[00:22:03] *** staylor has quit IRC[00:26:11] *** bookmark has quit IRC[00:27:07] *** bb010g has quit IRC[00:29:53] *** PasNox has quit IRC[00:32:47] *** Life-Blood has quit IRC[00:33:04] *** shintah has quit IRC[00:34:08] *** Codex_ has quit IRC[00:36:32] *** garFF has quit IRC[00:42:33] *** b4b has joined ##OpenGL[00:42:49] <b4b> what yall think of yosemite os?[00:43:04] <b4b> seems like win7 glass theme[00:44:26] <foobaz> i’ve heard it’s buggy, i’m sticking with 10.9 for now[00:45:06] <slime> it's pretty and functional[00:45:35] <foobaz> it’s a shame it’s still opengl 4.1[00:45:40] <slime> yeah[00:45:57] <foobaz> although that’s not so bad, 4.1 is very capable[00:46:04] <foobaz> much better than the years of being stuck on 2.1[00:46:10] *** tcsc has joined ##OpenGL[00:49:40] *** Barrett__ has quit IRC[00:50:54] *** Barrett_ has joined ##OpenGL[00:52:15] *** urraka has joined ##OpenGL[00:52:32] *** paul424 has quit IRC[00:55:04] *** pazul has joined ##OpenGL[00:57:56] *** goliath has quit IRC[00:59:38] *** bkre_ has quit IRC[01:04:04] *** paperMogwai has joined ##OpenGL[01:07:47] *** centrinia has joined ##OpenGL[01:09:34] *** Black0range has quit IRC[01:10:55] *** aethersis has quit IRC[01:11:09] *** Ad1 has quit IRC[01:16:34] *** BitPuffin has quit IRC[01:18:24] *** Waynes1 has quit IRC[01:19:06] *** hexagoxel_ has joined ##OpenGL[01:20:54] *** hexagoxel_ has quit IRC[01:21:07] *** hexagoxel_ has joined ##OpenGL[01:21:34] *** hexagoxel has quit IRC[01:21:42] *** hexagoxel_ has quit IRC[01:21:54] *** hexagoxel has joined ##OpenGL[01:30:04] *** indefini has quit IRC[01:33:40] *** paperMogwai has quit IRC[01:36:53] *** Crehl has quit IRC[01:38:54] *** hexagoxel has quit IRC[01:44:51] *** LifeBlood has joined ##OpenGL[01:48:04] *** derhass has quit IRC[01:52:19] *** bjz_ has quit IRC[01:53:51] *** HuntsMan has quit IRC[01:55:45] *** konom has quit IRC[02:00:52] *** hahuang61 has quit IRC[02:01:01] *** sparky has joined ##OpenGL[02:02:13] <sparky> glYoda: did you know amd doesn't allow default arguments to functions in glsl?[02:02:18] <sparky> nvidia does[02:02:27] *** konom has joined ##OpenGL[02:02:31] <sparky> like if you're passing bools for instance to a function[02:02:38] <sparky> it won't let you use bBlaBla=false)[02:02:39] *** SugarCube has joined ##OpenGL[02:03:04] <Stragus> I don't think that's part of the GLSL specs, so don't do that even if some drivers accept it[02:03:23] <Orion]> if(amd) {exit}[02:03:34] <sparky> ha ha[02:04:12] <sparky> glsl was designed to make things as difficult as possible[02:04:41] *** Twinklebear has joined ##OpenGL[02:05:01] <foobaz> AMD follows the GLSL spec, nvidia adds proprietary extensions[02:05:16] <foobaz> nvidia is basically trying to pull an internet explorer with “embrace and extend"[02:07:43] *** paperMogwai has joined ##OpenGL[02:07:50] <Stragus> Yet somehow Web page transition effects from IE 3 don't seem to have caught on[02:08:19] *** snakenerd has joined ##OpenGL[02:08:58] *** RyanPridgeon has quit IRC[02:09:33] *** snakenerd has quit IRC[02:10:21] *** RyanPridgeon has joined ##OpenGL[02:13:39] *** TyrfingMjolnir has quit IRC[02:13:56] <sparky> foobaz: perhaps but up until recently amd's compiler wouldn't even allow me to do stuff like if((g_iDstFlags&MY_POUND_DEFINE)!=0)[02:14:06] <sparky> I had to use explicit hex values[02:14:12] <sparky> now how messed up is that?[02:14:30] <foobaz> maybe it was an int/uint type mismatch[02:14:38] <sparky> definitely not[02:14:49] <foobaz> old versions of GLSL didn’t do implicit type conversions[02:14:51] <sparky> hmm[02:14:54] <foobaz> i think that was added in 3.3[02:14:55] <sparky> oh I see what you mean[02:15:24] <sparky> well I had this at the top[02:15:24] <sparky> #version 150[02:15:36] <sparky> you can still trigger the issue today in gsa[02:15:44] <sparky> gpu shader analyzer[02:15:53] <sparky> since they didn't update it since mid 2012[02:15:59] <foobaz> is g_iDstFlags an int or a uint?[02:16:07] <sparky> int[02:16:20] <foobaz> wait, #version 150? did that even have bitwise ops?[02:16:28] <foobaz> and if it did, were you allowed to use them on its?[02:16:33] <foobaz> on ints i mean[02:16:53] <foobaz> i’m not convinced that code is standard-compliant[02:17:02] <foobaz> we need a glsllint.com[02:17:51] *** pazul has quit IRC[02:18:38] <sparky> no not uint related[02:18:39] <sparky> just checked[02:18:44] <sparky> this is the error message it gives[02:18:44] <sparky> ERROR: 0:912: error(#143) Undeclared identifier undefined ERROR: 0:912: error(#132) Syntax error: '<' parse error ERROR: error(#273) 2 compilation errors. No code generated[02:18:57] <sparky> until I change the pounddefine to an explicit hex value[02:21:06] <sparky> strange[02:21:08] <sparky> even this fails[02:21:09] <sparky> int iFlag = VERTICAL_CUBE_MAP;[02:21:15] <sparky> same error message[02:21:38] <foobaz> maybe the AND produces an unsigned result, so you must compare it against another unsigned[02:21:46] <sparky> it's just declared as #define MY_VALUE 0x100[02:21:52] <foobaz> try 0u[02:21:55] <sparky> I'm using same name of course[02:21:59] <sparky> in the &[02:22:03] <sparky> I get that I swapped name here[02:22:06] <sparky> :)[02:22:09] <sparky> but it's matching[02:22:10] <foobaz> 0u would also be unsigned[02:22:23] *** SleekoNiko has joined ##OpenGL[02:24:38] <sparky> looks like they just didn't support pound defines period[02:24:41] <sparky> up until fairly recent[02:26:41] <slime> foobaz: there's https://www.khronos.org/opengles/sdk/tools/Reference-Compiler/[02:27:01] <foobaz> ah nice, now someone just has to wrap it in a web interface[02:27:07] <slime> amd's compiler has always supported #define[02:27:24] <slime> maybe there have been bugs in its implementation though, I dunno[02:28:14] <foobaz> looks like #define has been in GLSL since the beginning[02:30:33] *** paperManu_ has quit IRC[02:30:37] *** paperMogwai has quit IRC[02:31:20] *** paperManu_ has joined ##OpenGL[02:33:46] *** Demon_Fox has quit IRC[02:36:40] *** paperMogwai has joined ##OpenGL[02:39:56] *** Zupoman has quit IRC[02:44:57] *** puerum has joined ##OpenGL[02:49:17] *** centrinia has quit IRC[02:55:23] *** meoblast001 has joined ##OpenGL[02:55:25] *** AbigailBuccaneer has quit IRC[02:55:45] *** AbiBuccaneer has joined ##OpenGL[02:57:41] *** AbigailBuccaneer has joined ##OpenGL[02:58:46] *** AbigailBuccaneer has left ##OpenGL[03:05:07] *** YarnFive8 has joined ##OpenGL[03:09:20] *** Orion] has quit IRC[03:10:38] *** SleekoNiko has quit IRC[03:10:56] *** SleekoNiko has joined ##OpenGL[03:15:45] *** iKlsR has joined ##OpenGL[03:17:29] *** luxigo has joined ##OpenGL[03:19:06] *** rosseaux has quit IRC[03:22:27] *** rosseaux has joined ##OpenGL[03:22:42] <sparky> foobaz: though you can still see the lack of pounddefines being supported in gpu shader analyzer it was also true for amd drivers about 6 months ago[03:22:54] *** TyrfingMjolnir has joined ##OpenGL[03:24:34] *** luxigo has quit IRC[03:24:48] *** rosseaux has quit IRC[03:25:42] <foobaz> i have a hard time believing AMD was missing such a basic feature[03:26:16] <foobaz> i think it’s more likely that you had an type mismatch bug and the compiler gave you an unhelpful error message[03:26:45] <sparky> okay :)[03:28:15] *** rosseaux has joined ##OpenGL[03:34:37] *** Demon_Fox has joined ##OpenGL[03:35:15] <sparky> foobaz: yeah ok whenever I use hex that's where it goes wrong[03:35:18] <sparky> this compiles[03:35:20] <slime> yep, #defines have always worked on AMD/ATI drivers (or at least, after writing GLSL code that has #defines that multiple millions of PC users have run, I haven't got a bug report indicating that #defines in shaders are an issue)[03:35:21] <sparky> #define FUNKYTOWN 100[03:35:24] <sparky> and this does not[03:35:26] <sparky> #define FUNKYTOWN 0x100[03:35:36] <sparky> when I try and pass it to a variable that is[03:35:37] <foobaz> so obviously #define works[03:35:43] <sparky> doesn't matter if it's uint or int[03:35:57] <sparky> whenever I'm using 0x for the value I can't pass it on[03:36:05] <sparky> but on todays drivers from AMD I can[03:36:08] <sparky> six months ago I couldn't[03:36:21] <sparky> and in gsa I still can't since it's old[03:36:31] <sparky> is there a particular way I should be declaring hex in glsl[03:36:37] <foobaz> the GLSL 3.3 spec does not contain the strings “hex” or “0x”[03:36:43] <foobaz> i don’t think it supports it[03:36:45] <slime> glsl 1.50 does[03:36:55] <foobaz> oh wait sorry[03:36:58] <sparky> well I ddi have #version 150[03:37:04] <sparky> they must have left it out by mistake then[03:37:10] <foobaz> it does, it is in section 4 variables and types[03:37:22] <foobaz> specifically 4.1.3 integers[03:37:35] <foobaz> what is the line where you use the #define?[03:37:45] <sparky> hmm?[03:37:52] *** LifeBlood has quit IRC[03:38:31] <foobaz> maybe the bug is on the line where you use the #define[03:38:39] <foobaz> instead of on the line with the #define statement[03:38:47] <sparky> http://www.heypasteit.com/clip/1NSR[03:38:55] <sparky> if I remove the 0x then it compiles[03:39:11] <sparky> and the error is[03:39:12] <sparky> ERROR: 0:15: error(#143) Undeclared identifier undefined ERROR: 0:15: error(#132) Syntax error: '<' parse error ERROR: error(#273) 2 compilation errors. No code generated[03:39:25] <sparky> I've tried passing the pounddefine to both an int and to an uint[03:39:28] <sparky> either way fails[03:39:39] *** paperMogwai has quit IRC[03:40:46] <foobaz> fwiw the spec says 0x100 is an int, not a uint[03:40:57] <sparky> np, one sec[03:41:19] *** shahanm has joined ##OpenGL[03:41:46] <sparky> http://www.heypasteit.com/clip/1NSS[03:41:58] <sparky> RROR: 0:15: error(#143) Undeclared identifier undefined ERROR: 0:15: error(#132) Syntax error: '<' parse error ERROR: error(#273) 2 compilation errors. No code generated[03:43:21] <sparky> and again removing the 0x results in instant success[03:45:07] *** bb010g has joined ##OpenGL[03:45:25] *** TyrfingMjolnir has quit IRC[03:45:45] <sparky> he he, this doesn't work either #define FUNKYTOWN ( int( 0x100 ) )[03:45:56] <foobaz> yea that’s weird[03:46:00] <sparky> something in there is hellbent on not providing hex[03:48:20] *** LifeBlood has joined ##OpenGL[03:53:30] *** Jaus has quit IRC[03:54:42] *** iKlsR has quit IRC[04:03:57] *** shingshang has joined ##OpenGL[04:06:34] *** robot-beethoven has joined ##OpenGL[04:19:37] *** RyanPridgeon has quit IRC[04:24:19] *** Life-Blood has joined ##OpenGL[04:24:41] *** LifeBlood has quit IRC[04:26:41] *** pizthewiz has quit IRC[04:28:11] *** stoopdapoop has joined ##OpenGL[04:29:41] *** det has quit IRC[04:30:12] *** det has joined ##OpenGL[04:33:48] *** jbud has joined ##OpenGL[04:33:51] *** TyrfingMjolnir has joined ##OpenGL[04:38:26] *** TyrfingMjolnir has quit IRC[04:38:27] *** paperManu_ has quit IRC[04:39:21] *** kaleid0 has quit IRC[04:46:32] *** luxigo has joined ##OpenGL[04:52:56] *** samrat has joined ##OpenGL[04:53:24] *** xissburg has quit IRC[04:54:28] *** DolpheenDream has joined ##OpenGL[04:55:16] *** Cabanossi has quit IRC[04:57:40] *** Cabanossi has joined ##OpenGL[05:06:56] *** b4b has quit IRC[05:09:22] *** kapil_ has joined ##OpenGL[05:10:28] *** TunnelCat has quit IRC[05:15:05] *** centrinia has joined ##OpenGL[05:33:02] *** zajfy has quit IRC[05:50:04] *** qeed has quit IRC[05:50:09] *** samrat has quit IRC[05:52:57] *** tcsc has quit IRC[06:00:52] *** b4b has joined ##OpenGL[06:04:17] *** b4b has quit IRC[06:07:26] *** meoblast001 has quit IRC[06:10:31] *** TyrfingMjolnir has joined ##OpenGL[06:11:53] *** Life-Blood has quit IRC[06:22:46] *** JoshSyn has joined ##OpenGL[06:28:52] *** luxigo has quit IRC[06:34:37] *** SnakeTrailSteve has quit IRC[06:35:01] *** luxigo has joined ##OpenGL[06:36:17] *** shahanm has quit IRC[06:36:26] *** LifeBlood has joined ##OpenGL[06:39:43] *** Zupo has joined ##OpenGL[06:42:09] *** SnakeTrailSteve has joined ##OpenGL[06:43:06] *** samrat has joined ##OpenGL[06:43:18] *** Zupo2 has quit IRC[06:49:46] *** samrat has quit IRC[06:51:57] *** BreadProduct has quit IRC[06:53:21] *** shahanm has joined ##OpenGL[06:53:23] *** BreadProduct has joined ##OpenGL[06:54:00] *** samrat has joined ##OpenGL[06:58:17] *** samrat has quit IRC[06:58:23] *** CainJacobi has quit IRC[06:59:50] *** foreignFunction has joined ##OpenGL[07:01:03] *** samrat has joined ##OpenGL[07:02:11] *** foreignFunction has quit IRC[07:02:23] *** foreignFunction has joined ##OpenGL[07:07:51] *** samrat has quit IRC[07:09:06] *** samrat has joined ##OpenGL[07:11:01] *** Life-Blood has joined ##OpenGL[07:11:06] *** LifeBlood has quit IRC[07:13:51] *** samrat has quit IRC[07:15:15] *** samrat has joined ##OpenGL[07:25:20] *** indefini has joined ##OpenGL[07:31:38] *** Zupo2 has joined ##OpenGL[07:32:11] *** samrat has quit IRC[07:32:34] *** foreignFunction has quit IRC[07:34:06] *** Zupo has quit IRC[07:34:25] *** samrat has joined ##OpenGL[07:35:41] *** shintah has joined ##OpenGL[07:39:34] *** SnakeTrailSteve has quit IRC[07:41:11] *** hahuang61 has joined ##OpenGL[07:41:15] <glYoda> hey sparky[07:46:06] *** samrat has quit IRC[07:46:12] *** telex has quit IRC[07:48:40] *** telex has joined ##OpenGL[07:48:51] *** samrat has joined ##OpenGL[07:50:34] *** linuxuz3r has quit IRC[07:53:01] *** Gamecubic has quit IRC[07:57:01] *** neure has joined ##OpenGL[07:57:50] *** samrat has quit IRC[07:59:18] *** samrat has joined ##OpenGL[08:05:56] *** Zerflag has joined ##OpenGL[08:06:14] *** Twinklebear has quit IRC[08:09:19] *** Burga has joined ##OpenGL[08:10:10] *** jbud has quit IRC[08:11:28] *** sparky has quit IRC[08:12:33] *** urraka has quit IRC[08:13:09] *** MLM has quit IRC[08:15:53] *** samrat has quit IRC[08:16:50] *** clauswitt has joined ##OpenGL[08:16:53] *** Zupo has joined ##OpenGL[08:19:37] *** Zupo2 has quit IRC[08:20:13] *** BreadProduct has quit IRC[08:27:18] *** luxigo has quit IRC[08:29:08] *** samrat has joined ##OpenGL[08:29:24] *** Zupo2 has joined ##OpenGL[08:31:22] *** Beetny has joined ##OpenGL[08:31:40] *** luxigo has joined ##OpenGL[08:32:53] *** Zupo has quit IRC[08:38:36] *** Lemml has joined ##OpenGL[08:39:35] *** Jack has joined ##OpenGL[08:39:59] *** Jack is now known as Guest29953[08:40:17] *** roboman2444 has quit IRC[08:48:52] *** roboman2444 has joined ##OpenGL[08:51:56] *** PasNox has joined ##OpenGL[08:53:44] *** ibqn has joined ##OpenGL[08:57:07] *** abs25 has joined ##OpenGL[09:02:12] *** sricharanized has joined ##OpenGL[09:03:39] *** DolpheenDream has quit IRC[09:07:57] *** SleekoNiko has quit IRC[09:08:51] *** sricharanized has left ##OpenGL[09:09:29] *** Life-Blood has quit IRC[09:10:37] *** PasNox has quit IRC[09:13:25] *** BzztPloink has joined ##OpenGL[09:13:50] *** karab44 has joined ##OpenGL[09:14:57] *** TyrfingMjolnir has quit IRC[09:21:19] *** BadMrsFrosty has joined ##OpenGL[09:31:23] *** centrinia has quit IRC[09:32:24] *** garFF has joined ##OpenGL[09:33:34] *** centrinia has joined ##OpenGL[09:34:12] *** PasNox has joined ##OpenGL[09:37:13] *** kedgen has joined ##OpenGL[09:37:41] <kedgen> I'm trying to use cmake in the command line to build the examples found here: http://www.opengl-tutorial.org/download/[09:38:16] <kedgen> When I use glxinfo it says opengl version 2.1[09:38:45] <kedgen> when I use the application OpenGL Extension viewer is says 4.1. What gives?[09:39:36] *** navitsu has joined ##OpenGL[09:40:59] *** navitsu has quit IRC[09:41:36] *** ^esc2 has joined ##OpenGL[09:41:52] <slime> kedgen: what OS?[09:41:56] <Bloodust> what gpu?[09:41:58] *** ^esc2 is now known as navitsu[09:42:20] *** shahanm has quit IRC[09:46:28] *** CrizMobius has joined ##OpenGL[09:46:38] <slime> kedgen: if you're using Mac OS X, then X11 isn't native to the OS, and glxinfo uses GLX which is an interface between X11 and opengl[09:46:50] <slime> so you shouldn't use glxinfo (or X11 in general) on OSX[09:48:17] <slime> Mac OS X supports Core Profile GL 4.1 or 3.3 depending on the GPU, and legacy GL 2.1 as well[09:48:46] <BtbN> OSX has an amazing software renderer, which supports OpenGL4[09:48:56] <BtbN> and performs surprisingly well[09:49:20] <slime> for Core Profile GL3+ with GLFW in particular (which that tutorial uses), I think you need to use glfwWindowHint to request a forward-compatible GL3.2 context too[09:49:25] *** staylor has joined ##OpenGL[09:49:51] *** det has quit IRC[09:50:22] *** puerum has quit IRC[09:54:41] *** BadMrsFrosty is now known as L-Spiro[09:55:46] *** sricharanized has joined ##OpenGL[09:57:57] *** Jeanne-Kamikaze has joined ##OpenGL[09:58:23] *** sricharanized has left ##OpenGL[09:59:14] *** Keniyal has joined ##OpenGL[10:01:02] *** garFF has quit IRC[10:02:37] <kedgen> oh, just saw this! Thanks for the advice. My gpu is an NVidia GeForce GFX 680MX on Yosemite[10:03:29] <kedgen> when I tried to build the examples from the url mentioned above with cmake, it gave me an error saying that I'm using opengl 2.1[10:05:12] <slime> yeah, you might need to modify the source code to make it call glfwWindowHint(GLFW_OPENGL_FORWARD_COMPAT, 1);[10:05:48] <L-Spiro> Shouldn’t ##OpenGL redirect to ##Direct3D?[10:05:56] <slime> why's that?[10:06:12] <L-Spiro> To avoid state machines etc.[10:06:34] <slime> kedgen: you might also need to change the requested GL version to 3.2 for GLFW to work, not sure (although it will still create a GL4.1 context in the end)[10:06:56] <Adrinael> Nice troll attempt, 1/10[10:08:24] <L-Spiro> Come on Adrinael it was at least 1.5 out of 10.[10:08:40] *** Zupo3 has joined ##OpenGL[10:09:48] *** karab44 has quit IRC[10:11:07] *** Zupo2 has quit IRC[10:11:28] *** karab44 has joined ##OpenGL[10:21:23] *** kapil_ has quit IRC[10:22:17] *** bjz has joined ##OpenGL[10:22:27] *** redeemed has joined ##OpenGL[10:23:55] *** Khlorghaal has quit IRC[10:24:18] *** gareppa has joined ##OpenGL[10:26:54] *** karab44 has quit IRC[10:27:52] *** Folkol has joined ##OpenGL[10:29:49] *** goliath has joined ##OpenGL[10:30:02] *** ibqn has quit IRC[10:31:01] *** hexagoxel has joined ##OpenGL[10:33:57] *** glYoda has quit IRC[10:35:49] *** karab44 has joined ##OpenGL[10:36:07] *** luxigo has quit IRC[10:38:33] *** karab44 has quit IRC[10:45:30] *** hexagoxel has quit IRC[10:46:32] *** PasNox has quit IRC[10:46:43] <L-Spiro> Admin me.[10:48:36] *** gareppa has quit IRC[10:51:07] <L-Spiro> what the hell it’s been 5 minutes since I demanded admin and I am not admin yet. How long does it take?[10:51:52] <Bloodust> orange[10:52:06] <L-Spiro> I see your point.[10:52:51] *** luxigo has joined ##OpenGL[10:54:11] *** adsc has joined ##OpenGL[10:56:40] <L-Spiro> ::glGetIntegerv( GL_L_SPIRO_ADMIN_STATUS, &iRet ); // Returns false. Program can’t continue.[10:56:49] <L-Spiro> How do I fix this?[10:57:07] <slime> delete system32[10:57:19] <L-Spiro> No I did that yesterday and it is still here.[10:57:33] <slime> defenestrate GPU[10:57:38] <Bloodust> you can delete system32 on a monday[10:57:43] <Bloodust> cant*[10:57:51] <Bloodust> you can only do that on sunday evenings[10:58:15] *** ablepharus has joined ##OpenGL[10:58:25] <L-Spiro> Fudge pickles.[10:58:59] <L-Spiro> I can’t make the greatest games you will never play until this is fixed.[11:01:29] *** robot-beethoven has quit IRC[11:05:48] *** roxlu has joined ##OpenGL[11:06:17] *** kapil_ has joined ##OpenGL[11:07:06] *** Zupo3 has quit IRC[11:08:05] *** Crehl has joined ##OpenGL[11:08:29] <roxlu> Hi, I’ve been tinkering about how I can get the smoothest (2D) animation. i.e. I have a 2D plane that I want to move from top to bottom using some tween function. Because framerate is never 100% stable using time as the basis for the timings can result in irregular steps/translations. Though when I use the frame count as time basis the “delta time” steps AND they visuals will be smooth, but then it could be that a translation takes[11:08:29] <roxlu> longer then I want (due to the fact that the fps is not stable). I was wondering if someone here has done some tests with this and what they prefer?[11:09:16] *** PasNox has joined ##OpenGL[11:09:43] *** Zerflag has quit IRC[11:09:59] *** HuntsMan has joined ##OpenGL[11:10:38] *** mat^2 has quit IRC[11:11:00] *** mat^2 has joined ##OpenGL[11:11:39] <dahlia> roxlu: usually when the frame begins I measure time since the keyframe began and use that to interpolate between start and end position[11:11:57] <roxlu> dahlia: so using system time?[11:12:18] <dahlia> whatever is available similar to gettimeofday()[11:12:24] <roxlu> yeah exactly[11:12:33] <dahlia> which seems to have good enough accuracy and precision[11:12:56] <roxlu> have you noticed that the animation is suuuper smooth? Because when you use time the ‘dt’ won’t be the same for each frame[11:12:56] <dahlia> I think SDL has some timer you can call[11:13:07] <dahlia> ya its smooth[11:13:08] <roxlu> yeah I have my own high res timer[11:13:48] *** foobaz has quit IRC[11:15:21] <dahlia> roxlu: is this smooth enough? http://dahliaisland.wwweb3d.net:8099/screenshots/animTest06.wmv[11:16:18] <roxlu> Ok, I’m testing with both time and frame based. But I’m wondering if it actually makes any difference. Lets say I use frame based and one frame takes 14ms the other 16ms then delta time I feed into the tween function will be the same, 1 (due to the fact I’m using frame count as time bases and the frame is incremented with exactly 1). This will be 1 for -all- of the frames. Though when I use system time as basis and one frame takes 14[11:16:18] <roxlu> and the other 16ms, that would mean the delta time is different and the result of the tween not “smooth”.[11:16:38] * roxlu not sure if I can play wmv[11:16:42] <L-Spiro> #1: Delta time must always be measured in microseconds, not milliseconds.[11:16:52] *** Codex_ has joined ##OpenGL[11:16:59] <roxlu> L-Spiro: yep, I was doing that, but that doesn’t matter in this case[11:17:07] <dahlia> milliseconds is fine. Clocks arent really all that accurate[11:17:10] <roxlu> or even ns doesn’t matter[11:17:26] <L-Spiro> #2: Timers are to be avoided at all costs. They are not reliable, they don’t work, ignore them. Make your own. Better than using an OS object that needs to be awakened unreliable.[11:17:52] *** alllex229 has joined ##OpenGL[11:17:53] <roxlu> L-Spiro: so you would recommend using e.g. frame count as time basis?[11:17:57] *** AbiBuccaneer has quit IRC[11:17:58] <L-Spiro> dahlia, milliseconds are not fine.[11:18:00] <dahlia> I think m$oft specifies 5 ms tolerance on the system clock[11:18:16] <roxlu> L-Spiro: I think using frame count results in smoother animations[11:18:21] <L-Spiro> No roxlu, that is the absolute last thing I would recommend.[11:18:39] <dahlia> L-Spiro: references?[11:18:45] *** samrat has quit IRC[11:19:01] *** samrat has joined ##OpenGL[11:19:07] <L-Spiro> http://msdn.microsoft.com/en-us/library/windows/desktop/ms644904%28v=vs.85%29.aspx[11:19:38] <dahlia> references about milliseconds not being fine[11:19:45] <roxlu> lol[11:19:59] *** erhan_ has joined ##OpenGL[11:20:05] <roxlu> L-Spiro: why would that be the absolute last thing you would recommend?[11:20:08] *** slidercrank has quit IRC[11:20:12] <L-Spiro> dahlia, it will take me more than 3 seconds to dig up those.[11:20:33] <roxlu> dahlia: nice video :)[11:20:33] <L-Spiro> roxlu, because what is the speed of the computer on which your game runs?[11:20:37] <dahlia> well my animations look fine :)[11:20:37] *** hexagoxel has joined ##OpenGL[11:20:50] *** wutata- has joined ##OpenGL[11:20:51] <roxlu> dahlia: yeah[11:21:02] <roxlu> thanks for making that video btw![11:21:11] <L-Spiro> Making games based on FPS is why old games don’t run on current hardware. And how do you plan to address networking issues based on FPS?[11:21:13] <dahlia> yw[11:21:18] *** goliath has left ##OpenGL[11:21:37] <roxlu> L-Spiro: but that’s not what I was talking about[11:21:38] <slime> if your animation runs faster than normal on my 75Hz monitor, then you've done something wrong[11:22:21] <L-Spiro> slime, animations don’t “run”. The current point in any animation is just an interpolation between key frames.[11:22:41] <L-Spiro> There is no such thing as “speed” in a proper animation.[11:22:44] <slime> if the interpolation completes quicker than normal, then[11:22:53] <dahlia> I usually use VSYNC and measure time when the frame began and use that to imterpolate. You can't rely on frame rate being consistent anyway even if you use VSYNC[11:23:07] <dahlia> because sometimes it slows down[11:23:12] <L-Spiro> Then you are doing it wrong.[11:23:28] <dahlia> I'm doing what wrong?[11:23:37] <L-Spiro> (Yes, v-sync is not reliable.)[11:23:58] <L-Spiro> Using v-sync for anything other than its purpose is wrong.[11:23:59] *** karab44 has joined ##OpenGL[11:24:33] <dahlia> L-Spiro: am I using VSYNC wrong?[11:24:43] <L-Spiro> Depends.[11:24:45] * roxlu creating an image[11:24:53] <dahlia> depends on what?[11:24:58] <L-Spiro> Are you using it for any other reason beyond avoiding tearing?[11:25:11] <dahlia> yes[11:25:22] <L-Spiro> Then you are not using it for the correct reason.[11:25:36] <L-Spiro> V-sync exists to avoid tearing, not to help you time shit in your games.[11:25:50] <dahlia> I use it so I dont waste resources drawing 400 frames a second when the display only draws 60. Is that wrong?[11:26:02] <L-Spiro> V-sync for you is 60, for that other guy over there it is 75.[11:26:03] *** AbigailBuccaneer has joined ##OpenGL[11:26:45] <dahlia> I also use it so the computer fan runs quieter.[11:27:03] <L-Spiro> You don’t have to draw faster than the screen can present, but basing your game’s logical timers on v-sync is absolutely 100% an error.[11:27:14] <slime> I generally enabled vsync in order to reduce CPU/GPU usage and heat[11:27:20] <slime> enable *[11:27:39] <dahlia> I dont have any timers on vsync. I said I measure the time when the frame started.[11:27:54] *** Crehl has quit IRC[11:28:21] * roxlu almost ready ^.^[11:28:47] <dahlia> anyway my animations are smooth and run at the rate they were designed to run at[11:28:51] <L-Spiro> dahlia, why do you measure time since the frame started?[11:29:15] <dahlia> L-Spiro: so I know where to interpolate in the animation[11:29:35] <roxlu> ok, so first we need to know that tweening functions often use a value between 0 and 1 as their input (called t). The ’t’ value is the percentage of the current time. So when a tween takes 1000ms, then 500ms would be t = 0.5.[11:29:38] <dahlia> how do you interpolate in your animations?[11:30:02] <L-Spiro> dahlia, that is a bit of an advanced subject, but I have written an article about it.[11:30:04] <L-Spiro> Hold on.[11:30:20] <roxlu> So this is a graph of based on time and frame count: http://i.imgur.com/Oa9HZxj.png[11:30:30] *** alllex229 has quit IRC[11:30:48] <dahlia> L-Spiro: well I've written a complete animation system so I guess I could probably handle the subject here since you are telling me I do it wrong[11:30:51] <roxlu> both are pretty smooth .. but it made me curious what would be better to use. If you see the frame based on you can see that it’s completely smooth[11:30:55] <L-Spiro> Game logic needs to happen at fixed intervals. But rendering at random intervals. I explain it here: http://lspiroengine.com/?p=378[11:31:12] <dahlia> L-Spiro: you're doing it wrong :P[11:31:43] <dahlia> you're doing it wrong because events can be delayed[11:32:11] <dahlia> if you dont account for that possibility, your simulation is inaccrate[11:32:21] <dahlia> *inaccurate[11:33:15] * roxlu funny how my question is taking it’s own course :)[11:33:43] <Bloodust> roxlu L-Spiro is a troll so he dont care[11:34:01] <roxlu> hehe[11:34:16] <L-Spiro> Bloodust, I was trolling. I am not normally a troll.[11:34:50] <Bloodust> it'll take time until you build credibility here if you start by trolling[11:34:55] <Codex_> you cant anyway change how people write their game timing[11:34:59] <roxlu> I think the difference is between implicit and explicit functions, but please correct me if I’m wrong. I would say implicit functions won’t result in the smoothest reslut when using time as basis[11:35:11] <L-Spiro> dahlia, events get delayed. My full system accounts for it. But it takes a very very long explanation, and I have explained it many times.[11:35:24] *** foobaz has joined ##OpenGL[11:36:18] <dahlia> L-Spiro: I dont believe you and at this point I really dont care even if you do post a like to your article wich is obviously way to complex for my poor little brain to comprehend[11:36:41] <dahlia> *link[11:37:51] <L-Spiro> http://www.gamedev.net/topic/650640-is-using-static-variables-for-input-engine-evil/#entry5113267[11:38:18] <L-Spiro> The post just below it confirms.[11:38:28] <roxlu> Does someone knows any channels where developers from big game companies hang out?[11:38:41] <dahlia> sorry not going to waste my time even clicking it[11:38:42] <Codex_> your credibility looks like this: http://www.dilbert.com/strips/comic/2014-11-25/[11:38:43] <Bloodust> general programming channels I suppose[11:39:00] *** timsche has joined ##OpenGL[11:39:01] <AbigailBuccaneer> roxlu, ##OpenGL ;)[11:39:03] <roxlu> Bloodust: ok, like #programming, #gamedev[11:39:16] <Bloodust> let me rephrase[11:39:17] <L-Spiro> roxlu, I am the senior graphics programmer at Square Enix.[11:39:19] <slime> most serious developers don't hang out in general gamedev channels[11:39:22] <roxlu> hehe then I would expect more from pro game devs ^.^ (not to blame you guys, you’re great)[11:39:26] <Bloodust> more like language or API specific channels[11:39:34] <Bloodust> such as this, ##c++ ##c etc.[11:39:52] <AbigailBuccaneer> all us AAA gamedevs hang out in ##PHP, as a joke[11:39:56] <Bloodust> L-Spiro sure you are[11:40:03] <L-Spiro> Do you want a developer larger than Square Enix? Or would you just prefer me to be more professional?[11:40:11] <roxlu> ok, hehe from experience I know I better don’t ask this stuff in ##C or ##C++[11:40:24] *** abs25 has quit IRC[11:40:29] <Bloodust> right, thats because they are highly topical channels[11:40:48] <Bloodust> the topic is the language and nothing else[11:40:49] <roxlu> haha actually I would have expected someone who could give me a answer on this issue : )[11:40:54] <roxlu> yaeh[11:41:03] <L-Spiro> I trolled earlier. I do that. Blame me.[11:41:05] <roxlu> Bloodust: I’ll give #gamedev a try (fingerscrossed)[11:41:28] <L-Spiro> But you are asking about actual game development. My trolling steps aside.[11:41:39] <roxlu> hehe[11:41:44] <roxlu> my questions didn’t changed[11:42:12] <L-Spiro> I know.[11:42:55] <L-Spiro> Bloodust: https://www.youtube.com/user/LSpiro/about[11:44:23] <Bloodust> this prooves nothing and I dont need you to proove yourself to me[11:44:32] <L-Spiro> I don’t.[11:48:21] <jophish> AbigailBuccaneer: sup?[11:48:42] <L-Spiro> But if you are asking about no-delay input, I assure you I am the definitive source. As others would agree: http://www.gamedev.net/topic/652439-3d-game-engine-questions/#entry5125111[11:50:02] <Codex_> http://sivut.koti.soon.fi/~terop/face_shadow.png[11:50:25] <Bloodust> you keep repasting that link for some reason[11:50:44] <Codex_> bloodust: its so great[11:50:53] *** mat^2 has quit IRC[11:51:57] *** mat^2 has joined ##OpenGL[11:54:19] *** konom has quit IRC[11:56:40] *** kedgen has quit IRC[11:57:10] *** TheBix has joined ##OpenGL[11:58:33] *** garFF has joined ##OpenGL[12:05:41] *** clauswitt has quit IRC[12:15:45] *** paul424 has joined ##OpenGL[12:19:58] *** adsc has quit IRC[12:20:37] *** PasNox has quit IRC[12:21:32] *** RyanPridgeon has joined ##OpenGL[12:23:54] *** adsc has joined ##OpenGL[12:24:38] *** luxigo has quit IRC[12:35:03] *** faty has joined ##OpenGL[12:36:47] *** pazul has joined ##OpenGL[12:40:02] *** zzuegg has quit IRC[12:41:46] *** zzuegg has joined ##OpenGL[12:53:04] *** indefini has quit IRC[12:54:02] *** karab44 has quit IRC[13:01:51] *** centrinia has quit IRC[13:05:58] *** Beetny has quit IRC[13:08:09] *** doev has joined ##OpenGL[13:13:01] *** nopcode has joined ##OpenGL[13:13:24] <nopcode> hey, i've built qt5 without EGL (-no-egl) and my resulting application still depends on libegl, which seems to be supplied by mesa on the build box[13:13:34] <nopcode> but when trying to run the app on another box, it wont find that library[13:13:37] <nopcode> any ideas?[13:14:20] *** Keniyal_ has joined ##OpenGL[13:15:54] *** Keniyal has quit IRC[13:17:41] *** navitsu has quit IRC[13:17:54] *** samrat has quit IRC[13:18:56] *** Tag_ has joined ##OpenGL[13:19:31] *** paperManu_ has joined ##OpenGL[13:25:10] *** xissburg has joined ##OpenGL[13:25:32] *** BitPuffin has joined ##OpenGL[13:26:29] *** Ad1_RN has joined ##OpenGL[13:27:49] *** RyanPridgeon has quit IRC[13:28:36] *** clauswitt has joined ##OpenGL[13:28:46] *** Jeanne-Kamikaze has quit IRC[13:30:19] *** Ad1_RN has quit IRC[13:30:28] *** Ad1_RnR has quit IRC[13:30:34] *** Ad1 has joined ##OpenGL[13:30:37] *** indefini has joined ##OpenGL[13:34:55] *** navitsu has joined ##OpenGL[13:39:37] *** navitsu has quit IRC[13:39:46] *** Keniyal has joined ##OpenGL[13:40:13] *** TyrfingMjolnir has joined ##OpenGL[13:41:09] *** Keniyal_ has quit IRC[13:41:30] *** PasNox has joined ##OpenGL[13:46:24] *** slidercrank has joined ##OpenGL[13:46:28] *** zzuegg has quit IRC[13:47:00] *** Burga has quit IRC[13:47:44] *** zzuegg has joined ##OpenGL[13:49:23] *** Madsy has quit IRC[13:50:53] *** Madsy has joined ##OpenGL[13:50:53] *** Madsy has joined ##OpenGL[13:53:26] *** DolpheenDream has joined ##OpenGL[13:55:52] *** bic has joined ##OpenGL[13:56:32] <bic> any good links on opengl es 2 state machine?[13:58:45] *** navitsu has joined ##OpenGL[14:00:27] *** Keniyal_ has joined ##OpenGL[14:03:40] *** Keniyal has quit IRC[14:06:13] *** navitsu has quit IRC[14:07:18] *** Keniyal has joined ##OpenGL[14:08:06] *** mat^2 has quit IRC[14:08:26] *** mat^2 has joined ##OpenGL[14:08:45] *** jdolan has quit IRC[14:09:41] *** paperManu_ has quit IRC[14:10:42] *** TyrfingMjolnir has quit IRC[14:10:51] *** Keniyal_ has quit IRC[14:10:58] *** TyrfingMjolnir has joined ##OpenGL[14:11:00] *** jdolan has joined ##OpenGL[14:13:52] *** mat^2 has quit IRC[14:15:06] *** urraka has joined ##OpenGL[14:17:05] *** Waynes has joined ##OpenGL[14:17:53] *** timsche has quit IRC[14:18:12] *** garFF has quit IRC[14:20:42] *** kuldeepdhaka has joined ##OpenGL[14:21:58] *** rishin458 has joined ##OpenGL[14:23:45] *** mat^2 has joined ##OpenGL[14:24:12] *** rishin458 has quit IRC[14:24:53] *** navitsu has joined ##OpenGL[14:29:30] *** garFF has joined ##OpenGL[14:29:39] *** TheBix has quit IRC[14:31:53] *** navitsu has quit IRC[14:38:05] *** samrat has joined ##OpenGL[14:40:45] *** shintahW2 has joined ##OpenGL[14:44:04] *** shintahW has quit IRC[14:44:04] *** shintahW2 is now known as shintahW[14:44:22] *** jdolan has quit IRC[14:45:50] *** ruuns has joined ##OpenGL[14:46:16] *** tehrain has joined ##OpenGL[14:46:26] *** garFF has quit IRC[14:46:50] *** tehrain has quit IRC[14:47:16] *** tehrain has joined ##OpenGL[14:48:00] *** snakenerd has joined ##OpenGL[14:49:37] *** faenil_ has joined ##OpenGL[14:49:50] *** navitsu has joined ##OpenGL[14:49:57] *** garFF has joined ##OpenGL[14:50:01] <faenil_> hello people :) quick question[14:50:40] <faenil_> could it be that using GL_SHORT instead of GL_FLOAT for vertex position is dramatically reducing my framerate?[14:50:54] <faenil_> or is there any other issue that could cause the slowdown?[14:51:09] <faenil_> my usecase is: drawing millions of GL_POINTS[14:51:43] <faenil_> I get 60fps (vsync-locked) when shooting 3M points[14:51:51] <faenil_> with GL_FLOAT[14:51:58] <faenil_> if I use GL_SHORT instead I get 17fps[14:52:29] <faenil_> if I shoot 30mill points, I get 40fps vs 3fps[14:52:59] <faenil_> that looks quite an extreme slowdown...I was looking for confirmations that this is only because of the short->float conversion that has to happen on GPU[14:53:01] <BtbN> "drawing millions of GL_POINTS" i'd say that is your problem[14:53:14] <faenil_> why?[14:53:31] <BtbN> No driver is optimized for that, performance results will be entirely random[14:53:57] <faenil_> mm ok, so you agree it could just be the SHORT->FLOAT conversion?[14:54:02] <faenil_> that kills performance?[14:54:14] <BtbN> Drawing millions of points is what kills any performance[14:54:21] <faenil_> no, maybe you didn't get me :)[14:54:32] <faenil_> with floats I get 40fps, with shorts 3fps[14:54:39] <faenil_> drawing the same amount of points[14:54:41] <BtbN> What are you even doing that you need/want to draw so many points?[14:55:05] <Codex_> it would only kill perf, if hw cannot support shorts directly[14:55:10] <faenil_> it's an algorithm I'm working on[14:55:28] <BtbN> if you want to draw raw points, not using OpenGL would most likely yield the best performance.[14:55:30] <faenil_> Codex_: exactly...how can I know that?[14:55:40] <BtbN> You can't. It's driver internals.[14:55:44] *** ManDay has joined ##OpenGL[14:55:51] <BtbN> Could be the exact opposite on some other machine[14:55:52] <Codex_> you can always measure the time :-)[14:56:04] <faenil_> BtbN: well it's point with advanced lighting effects, I need opengl :)[14:56:06] <Codex_> for all the platforms you support :-)[14:56:11] <faenil_> Codex_: heheh[14:56:22] <BtbN> As far as i know OpenGL can't put lightning on points.[14:56:26] <neure> faenil_, try half floats[14:56:34] <faenil_> I needed shorts to compress the VBO data, but if it yields such a consistent slowdown..[14:56:37] *** navitsu has quit IRC[14:56:42] <neure> half floats are smaller than floats :)[14:57:04] <faenil_> neure: though only 10bits mantissa..[14:57:21] <neure> so how would you have used shorts then?[14:57:34] <faenil_> mm?[14:57:46] <neure> short is 16 bits as well[14:57:57] <faenil_> yep, but it's int[14:58:01] <neure> so?[14:58:06] <neure> you would normalize it in the shader?[14:58:24] <BtbN> 6 bits more precison do matter.[14:58:26] <faenil_> yes, which converts it to a bigger float (I guess)[14:58:42] <Codex_> if gl_float works ok, why not use it?[14:58:49] <faenil_> Codex_: memory usage[14:58:57] <Codex_> buy more mem :-)[14:59:05] <faenil_> can't buy VRAM :D[14:59:12] <neure> sure you can[14:59:18] <faenil_> really? how :D[14:59:22] <neure> you just need to get a whole new GPU[14:59:54] <neure> what kind of points those are, anyway?[14:59:59] <faenil_> this is like "hey let's use javascript everyone, there will CPUs which will be able to handle Unity engine coded in JS one day" :D[15:00:18] <faenil_> it's point, gl_points[15:00:30] <neure> i meant the attribute values[15:00:41] <neure> are the any kind of floats within some range?[15:00:57] <faenil_> the vertex attributes? or what do you mean?[15:01:09] <neure> yes, i did mean vertex attribute values[15:01:11] <faenil_> ah, yes[15:01:21] <neure> what is the range?[15:01:34] <faenil_> they're natively floats, which I scale to be within GL_SHORTS range[15:02:06] <neure> you mean before you put them into buffer for GL to consume as GL_SHORTS?[15:02:18] <faenil_> yes[15:02:22] <neure> right[15:02:40] <neure> what do you use this for?[15:03:12] <faenil_> this what[15:03:18] <neure> never mind[15:03:41] <faenil_> it's an algorithm I'm coding, and the compression is to save data on vbo so that the same GPU can handle more points[15:03:52] <faenil_> nothing strange there ;)[15:04:17] <neure> so all you want to do is to compress?[15:04:33] <neure> you can compress, and it will cost some performance on some GPUs, that is pretty much the end result[15:04:52] <faenil_> yea, I didn't expect "some" to be a 10x slowdown[15:05:03] <neure> it is totally GPU specific[15:05:04] <faenil_> that's why I joined to ask if it was "expected"[15:05:14] <faenil_> okay ;)[15:05:14] <neure> some GPUs do attribute conversions with shader code[15:05:22] <neure> some GPUs have dedicated HW[15:05:30] <neure> both of those can become bottleneck[15:05:33] <faenil_> I'm using an AMD hd5850, fwiw[15:05:52] <neure> dedicated HW for attribute conversion is not so common any more[15:05:58] *** centrinia has joined ##OpenGL[15:06:09] <neure> and IIRC AMD does attribute fetching and conversion using shaders[15:06:14] <Codex_> faenil_: you have a screenshot of your millions of gl_points?[15:06:28] <faenil_> can't show it at the moment, sorry :/[15:06:53] <neure> you could try to read unnormalized ints (not shorts) in shader[15:07:09] <neure> split that to 16 bit parts and normalize with your own shader code[15:07:17] <neure> see if that has any different performance[15:07:33] <neure> weres your shorts well aligned?[15:07:42] <neure> or did you tightly pack 3 shorts for position?[15:07:44] <faenil_> you mean 32bit ints?[15:07:47] <neure> yes[15:07:58] <neure> you should make sure your attributes are nicely aligned[15:08:00] <faenil_> ah, right...let me check[15:08:29] <roxlu> Ok guys, I created 2 test applications, one which uses system time and the other frame count to do the tween interapolation I discussed a couple of hours ago. Anyone who wants to give it a try? I want to do a tiny experiment to see what version you like best[15:08:38] <faenil_> right...not aligned[15:08:40] <neure> HW tends to have caches here and there, and if your data doesnt fit to one cache line, you pay performance penalty[15:08:54] <faenil_> so I need 16bits more to align it right?[15:09:00] <faenil_> yeah..[15:09:02] <neure> possibly[15:09:14] <neure> i dont know AMD internals[15:09:15] <faenil_> I aligned it when I compressed colours, but forgot to do it for position as well[15:09:21] <Codex_> i got this with gl_points: http://sivut.koti.soon.fi/~terop/clouds2.png[15:09:30] *** bjz has quit IRC[15:09:41] <faenil_> Codex_: what is it? :)[15:09:54] <faenil_> I mean, the semantics[15:09:59] <neure> "Program" :D[15:10:01] <Bloodust> jesus[15:10:08] <faenil_> makes sense :D[15:10:09] <Bloodust> its the 2nd coming of jesus[15:10:10] <Codex_> its some sphere shapes[15:10:56] <Codex_> the main algorithm just rejects any gl_points that are not inside a sphere using x^2+y^2+z^2<r^2[15:11:00] <roxlu> Codex_, Bloodust, dahlia, L-Spiro wanna give it a try?[15:11:03] *** bjz has joined ##OpenGL[15:11:12] <Bloodust> hum[15:11:55] <Bloodust> roxlu where are they?[15:12:05] <roxlu> are you on mac or win?[15:12:08] <Bloodust> win[15:12:11] <L-Spiro> I didn’t see the question.[15:12:46] <faenil_> neure: BINGO! :)[15:12:52] <roxlu> L-Spiro: see a couple lines up[15:13:01] <faenil_> with 32bit alignment it goes from 10 to 60fps![15:13:37] <faenil_> woohoo...who would've thought it could make so much difference[15:13:56] <neure> nice[15:14:14] <faenil_> thanks for the hint, I did it for the colour, completely forgot to do it here as well :)[15:14:23] <neure> do you interleave position and color?[15:14:31] <faenil_> yep[15:14:42] <neure> make sure the whole thing is well aligned[15:14:52] *** navitsu has joined ##OpenGL[15:15:02] <Codex_> I keep position and colour in different arrays :-)[15:15:07] <faenil_> yeah the vertex data now is[15:15:09] <neure> well, at least that each attribute fetch, if you one can imagine such, is well aligned[15:15:16] <faenil_> 64bits pos + 32bits col + 32bits uv[15:15:51] <neure> thats sounds like a nice alignment of 128 bits[15:15:57] <faenil_> it's a pity to waste 16bits like this...but if it kills the perf :)[15:16:07] <faenil_> maybe I should fill it with some useful info :p[15:16:22] <neure> at least you have room for say extra uv or something[15:16:27] <faenil_> mm no, wouldn't help as it would still be another attrib[15:16:37] <neure> you can unpack it with a shader[15:16:44] <faenil_> doesn't it have to be aligned on attrib level?[15:16:53] <faenil_> right[15:16:59] <neure> treat it as vec4 or something[15:17:09] <faenil_> yeah[15:17:30] <neure> or raw ints and unpack & convert to float manually in the shader if you need more flexibility[15:17:33] <faenil_> good idea, I just need something useful to put in there :) but thanks people![15:17:40] <neure> no prob[15:18:17] *** karab44 has joined ##OpenGL[15:19:29] *** navitsu has quit IRC[15:21:41] *** MissionCritical has quit IRC[15:23:56] *** Mission-Critical has joined ##OpenGL[15:25:56] *** bic has quit IRC[15:26:25] *** JoshSyn has quit IRC[15:27:22] *** jdolan has joined ##OpenGL[15:32:10] *** sparky has joined ##OpenGL[15:33:28] *** SnakeTrailSteve has joined ##OpenGL[15:38:29] *** Mission-Critical has quit IRC[15:38:47] *** zajfy has joined ##OpenGL[15:39:39] *** navitsu has joined ##OpenGL[15:45:06] *** Gamecubic has joined ##OpenGL[15:45:17] *** matpow2 has joined ##OpenGL[15:45:33] *** Tag_ has quit IRC[15:47:09] *** mat^2 has quit IRC[15:47:29] *** navitsu has quit IRC[15:55:35] *** kuldeepdhaka has quit IRC[15:55:45] *** _ami_ has joined ##OpenGL[15:56:09] <_ami_> Hello[15:56:18] <_ami_> I am new to opengl.[15:56:29] <Bloodust> ok[15:56:51] <_ami_> Which one should I learn first? 1. open gl or 2. opengl es?[15:57:17] <Bloodust> no big difference[15:57:24] <Bloodust> what system are you targeting?[15:57:36] <_ami_> I have great interest in graphics .. and i work for a semiconductor company.[15:57:57] <_ami_> Bloodust: mainly embedded.[15:59:09] <Bloodust> then you should probably start with opengl es[15:59:42] <_ami_> Bloodust: Please suggest a book.[15:59:56] <_ami_> for opengl es.[16:00:00] <Bloodust> I cant suggest a book for ES[16:00:35] <_ami_> link or sample? Can I code and test my apps on PC as well. do we have emulator for PC[16:00:57] <_ami_> say for Mali GPUs.[16:01:41] <Bloodust> if you have opengl ES 3.1 support then you can read desktop gl tutorials[16:02:27] <_ami_> which version of desktop gl? >3.3?[16:02:27] <Bloodust> latest opengl 4.5 and opengl es 3.1 are compatible with each other[16:02:36] <_ami_> ah! nice.[16:03:31] <_ami_> opengl es 2.0 == desktop gl <version?>[16:03:38] <Bloodust> and es 3.0 is compatible with desktop 4.3[16:03:49] <Bloodust> es 2.0 is based rougly on opengl 2.0[16:04:20] *** Zupo2 has joined ##OpenGL[16:04:25] <_ami_> umm.. our hw support es 3.0[16:04:52] *** navitsu has joined ##OpenGL[16:04:54] <Bloodust> thats good[16:04:57] <Bloodust> brb 15mins[16:04:59] <Bloodust> gotta go home[16:05:11] <_ami_> Bloodust: okay, thanks for your help.[16:05:25] <AbigailBuccaneer> ES 3.0 is compatible with desktop 4.3, but ES 3.0 and desktop 3.3 are roughly the same... i try and write my rendering code to target the intersection of ES3 and desktop 3.3[16:05:44] <AbigailBuccaneer> of course, it's somewhat tricky to know what that intersection is unless you have experience with both APIs individually[16:06:45] <AbigailBuccaneer> (when i say "roughly" the same i do mean quite rough indeed)[16:07:13] <_ami_> AbigailBuccaneer: hrm, porting ES 3.0 code on desktop would be a nightmare then. :)[16:07:48] *** kuldeepdhaka has joined ##OpenGL[16:08:04] *** jdolan has quit IRC[16:08:14] <AbigailBuccaneer> if you write ES3 code, then you can get it running on 4.3 with very little hassle - there are extensions (that i think became core in 4.3) that add ES-specific stuff, just to make this painless[16:08:29] *** jdolan has joined ##OpenGL[16:08:34] <AbigailBuccaneer> but if you write ES3 code then you'll find that you can get it running on 3.3 with few modifications too[16:09:17] *** clauswitt has quit IRC[16:09:22] *** DolpheenDream has quit IRC[16:09:57] *** centrinia has quit IRC[16:10:25] *** RockLee has joined ##OpenGL[16:10:26] *** goliath has joined ##OpenGL[16:12:13] *** navitsu has quit IRC[16:18:53] *** matpow3 has joined ##OpenGL[16:19:04] *** goliath has quit IRC[16:20:26] *** centrinia has joined ##OpenGL[16:21:30] *** shintahW2 has joined ##OpenGL[16:22:10] *** _ami_ has quit IRC[16:22:26] *** matpow2 has quit IRC[16:22:54] *** pazul has quit IRC[16:24:44] *** shintahW has quit IRC[16:24:44] *** shintahW2 is now known as shintahW[16:24:58] *** MissionCritical has joined ##OpenGL[16:26:08] *** garFF has quit IRC[16:29:07] *** MLM has joined ##OpenGL[16:29:52] *** navitsu has joined ##OpenGL[16:33:00] *** telex has quit IRC[16:34:13] *** garFF has joined ##OpenGL[16:34:39] *** telex has joined ##OpenGL[16:36:57] *** navitsu has quit IRC[16:38:22] *** RockLee has quit IRC[16:42:47] *** Folkol has quit IRC[16:46:23] *** det has joined ##OpenGL[16:48:42] *** kuldeepdhaka has quit IRC[16:48:58] *** sedeki has joined ##OpenGL[16:51:23] *** kapil_ has quit IRC[16:51:44] *** Zupo3 has joined ##OpenGL[16:52:35] *** HuntsMan has quit IRC[16:52:51] *** luxigo has joined ##OpenGL[16:54:02] *** Zupo2 has quit IRC[16:54:48] *** navitsu has joined ##OpenGL[16:57:45] *** sedeki has quit IRC[16:59:37] *** doev has quit IRC[17:01:41] *** navitsu has quit IRC[17:02:29] *** faenil_ has quit IRC[17:03:02] *** Keniyal has quit IRC[17:03:50] *** Codex_ has quit IRC[17:04:28] *** roboman2444 has quit IRC[17:05:01] *** kuldeepdhaka has joined ##OpenGL[17:09:49] *** sedeki has joined ##OpenGL[17:10:05] *** Codex_ has joined ##OpenGL[17:11:01] *** mat^2 has joined ##OpenGL[17:11:37] *** roboman2444 has joined ##OpenGL[17:12:42] *** matpow3 has quit IRC[17:17:27] *** delicado has joined ##OpenGL[17:18:20] *** Zerflag has joined ##OpenGL[17:19:53] *** navitsu has joined ##OpenGL[17:21:25] *** Waynes1 has joined ##OpenGL[17:22:54] *** Waynes has quit IRC[17:24:30] *** centrinia has quit IRC[17:24:33] *** redeemed has quit IRC[17:27:21] *** navitsu has quit IRC[17:36:22] *** delicado has quit IRC[17:39:14] *** wutata- has quit IRC[17:40:52] *** L-Spiro has quit IRC[17:43:51] *** ShadowIce has joined ##OpenGL[17:44:04] *** japro has joined ##OpenGL[17:44:52] *** navitsu has joined ##OpenGL[17:45:07] *** aphorisme has joined ##OpenGL[17:45:12] <japro> hmm, i guess carmacks reverse is faster in my case[17:45:39] <japro> also amusingly the difference is: comment out one line, add glDepthFunc(GL_GEQUAL)[17:45:58] <roboman2444> carmacks reverse is copyrighted, right?[17:46:02] <japro> patented[17:46:04] <roboman2444> by creative sound of all places[17:46:58] *** japro has quit IRC[17:47:01] *** neurre has joined ##OpenGL[17:47:23] *** seangrove has quit IRC[17:47:41] *** goliath has joined ##OpenGL[17:47:49] *** seangrove has joined ##OpenGL[17:47:56] *** delicado has joined ##OpenGL[17:47:59] <delicado> i know this has been ask many times, what's fastest to render on todays graphics accelerators, quads or triangles? i searched google about this but seen conflicting answers, some say triangles are rendered faster.. but i think quads should render faster, the reason is geometries composed of quads cost little in memory and so less memory fetch and less bandwidth is used[17:49:11] <goliath> nobody implements quad rendering in hardware[17:49:19] <AbigailBuccaneer> delicado, triangles. quads will be decomposed to triangles at some point[17:49:36] *** ShadowIce has quit IRC[17:50:20] <AbigailBuccaneer> if you're doing indexed rendering (ie. with a GL_ELEMENT_ARRAY_BUFFER) it won't be much more memory at all. triangles are more elegant - you can't specify a non-planar triangle, interpolating them is easier, etc. etc.[17:51:37] *** navitsu has quit IRC[17:52:22] <tehrain> aren't quads phased out anyway in modern OpenGL?[17:52:42] <goliath> yep[17:53:25] <delicado> okay thanks guys. i thought old knowledge is applicable on todays hardware, i feel stupid already :)[17:53:31] *** linuxuz3r has joined ##OpenGL[17:54:00] <Codex_> you can always mix quads and triangles. (push_back to different array will do it)[17:58:27] <Stragus> No, you have to decompose quads into triangles one way or the other[18:00:56] *** panda81_ has joined ##OpenGL[18:03:05] <tehrain> might as well just use triangles[18:03:32] <tehrain> you're using 4 vertices either way to make a quad, so just use the same 4 vertices to make 2 triangles with one shared side and you're good[18:03:52] *** doev has joined ##OpenGL[18:04:00] *** centrinia has joined ##OpenGL[18:09:20] *** konom has joined ##OpenGL[18:09:52] *** navitsu has joined ##OpenGL[18:13:58] *** BitPuffin has quit IRC[18:14:01] *** adsc has quit IRC[18:16:18] *** BearishMushroom has joined ##OpenGL[18:16:42] *** tehrain has quit IRC[18:17:16] *** garFF has quit IRC[18:17:17] *** navitsu has quit IRC[18:20:07] *** sparky has quit IRC[18:20:10] *** SnakeTrailSteve has quit IRC[18:21:59] *** CrizMobius has quit IRC[18:22:24] *** tehrain has joined ##OpenGL[18:22:48] *** stefkos has joined ##OpenGL[18:23:24] *** SnakeTrailSteve has joined ##OpenGL[18:23:37] *** BitPuffin has joined ##OpenGL[18:24:53] *** cr`nge has joined ##OpenGL[18:27:07] *** bb010g has quit IRC[18:27:07] *** PasNox has quit IRC[18:32:45] *** doev has quit IRC[18:33:26] *** xishem has quit IRC[18:34:52] *** navitsu has joined ##OpenGL[18:35:00] *** centrinia has quit IRC[18:36:03] *** erhan_ has quit IRC[18:38:07] *** Crehl has joined ##OpenGL[18:39:57] *** panda81_ has quit IRC[18:42:01] *** navitsu has quit IRC[18:43:13] *** staylor_ has joined ##OpenGL[18:46:47] *** aphorisme has quit IRC[18:47:53] *** aphorisme has joined ##OpenGL[18:50:20] *** hahuang61 has quit IRC[18:56:31] *** Match has joined ##OpenGL[18:59:49] *** navitsu has joined ##OpenGL[19:05:01] *** paul424 has quit IRC[19:06:45] *** navitsu has quit IRC[19:11:05] *** cr`nge has quit IRC[19:11:25] *** Biliogadafr has joined ##OpenGL[19:11:43] *** ManDay has quit IRC[19:13:03] *** BitPuffin has quit IRC[19:15:19] *** anivemin has joined ##OpenGL[19:21:30] *** SnakeTrailSteve has quit IRC[19:21:31] *** HuntsMan has joined ##OpenGL[19:22:45] *** ManDay has joined ##OpenGL[19:22:50] *** DolpheenDream has joined ##OpenGL[19:24:03] *** shingshang has quit IRC[19:24:42] *** cr`nge has joined ##OpenGL[19:24:52] *** navitsu has joined ##OpenGL[19:28:21] *** hahuang61 has joined ##OpenGL[19:28:48] *** Misu has joined ##OpenGL[19:31:57] *** navitsu has quit IRC[19:32:06] *** RyanPridgeon has joined ##OpenGL[19:32:50] *** Khlorghaal has joined ##OpenGL[19:33:15] *** SnakeTrailSteve has joined ##OpenGL[19:37:04] *** navitsu has joined ##OpenGL[19:38:52] *** neurre has quit IRC[19:39:15] <navitsu> test[19:40:31] <Codex_> case[19:42:48] *** Gamecubic has quit IRC[19:43:12] *** Gamecubic has joined ##OpenGL[19:43:52] *** aphorisme has quit IRC[19:43:59] *** navitsu has quit IRC[19:44:14] *** goliath has left ##OpenGL[19:44:16] *** navitsu has joined ##OpenGL[19:46:32] *** foreignFunction has joined ##OpenGL[19:50:18] *** Jeanne-Kamikaze has joined ##OpenGL[19:55:21] *** aphorisme has joined ##OpenGL[19:56:09] *** derhass has joined ##OpenGL[19:57:06] *** glYoda has joined ##OpenGL[19:57:28] <ManDay> please[19:59:20] *** Match has quit IRC[20:00:41] <Codex_> I had nice problem with sprites disappearing when I added some completely unrelated sprites to the scene :-)[20:01:04] <Codex_> it turned out to be std::map collision problem[20:01:18] <Bloodust> ha[20:04:50] *** aphorisme has quit IRC[20:10:06] *** hexagoxel has quit IRC[20:10:27] <ManDay> And as a consequence, you have decided to ditch C++ and henceforth use C[20:10:53] <ManDay> …and he lived happily ever after[20:11:03] <Codex_> hmm, nope, I just fixed the bug.[20:11:04] *** Zupoman has joined ##OpenGL[20:11:22] <Codex_> I hate huge rewrites.[20:12:39] <ManDay> It's just two characters less...[20:13:17] *** pazul has joined ##OpenGL[20:13:40] <Codex_> modifying existing code seems like quitting... You only do it when you no longer trust that it's perfect[20:14:29] *** Zerflag has quit IRC[20:16:02] *** Zupo3 has quit IRC[20:16:12] <Bloodust> you must feel like quitting every day[20:17:03] <derhass> I never _trust_ or even assume existing code to be perfect[20:17:09] <derhass> especially when I wrote it myself[20:17:19] <Codex_> naah, don't need to. I just try to write perfect code every time.[20:17:51] <Codex_> if you're not even trying to make it perfect, its another form of quitting[20:18:19] <Bloodust> Codex_ is that the reason why it took you two years to display even the most simplest graphics with opengl? :D[20:18:20] *** Demon_Fox has quit IRC[20:18:42] <Codex_> bloodust: well, experimenting with opengl features just takes time[20:21:16] <derhass> Codex_:that's all within the limits and constraints of reality[20:21:17] <Codex_> bloodust: but I wanted to fully utilize every feature I depend on. So one feature takes long time.[20:21:17] *** Zerflag has joined ##OpenGL[20:21:17] <delicado> is it possible to pack index data and vertex data into a single VBO? so that i only have a single VBO for a mesh.[20:21:17] <derhass> delicado: yeas, but not really advisable[20:21:36] <delicado> derhass: why that is, can it be rendered?[20:21:53] <derhass> sure, it can be rendered[20:22:06] <derhass> you might just work against the optimizations of your implementation[20:22:18] <derhass> although I'm not sure how relevant that is nowadays[20:22:57] <Codex_> delicado: I thought you'd use vao for that?[20:23:08] <delicado> well, i will just try i guess.[20:23:21] <delicado> Codex: i forgot whats buffers are called.[20:23:34] <delicado> but i do know how to use them[20:23:48] <ManDay> Codex_: I share your sentiment about trying to write perfect code and spending excessively long times on internalizing every smallest aspect about libs like OpenGL. But your assumption that your code *is* perfect is perfect bullshit[20:24:18] <ManDay> I go by the same ideas but whenever I look at my own code I wonder who wrote that bullshit[20:24:38] <ManDay> if ( var != true ) are just the more harmless gems I find in that crap[20:25:04] *** pizthewiz has joined ##OpenGL[20:25:18] *** Alina-malina has quit IRC[20:25:40] <Codex_> manday: well, you just need to decide what your definition of "perfect" is. Big problems comes if you can't decide that now it's perfect.[20:26:06] <Codex_> tweaking the code endlessly is another horror story[20:26:39] <Bloodust> please[20:26:41] *** Vtec234 has joined ##OpenGL[20:26:53] <Bloodust> my guess is at max 1% strives for perfection[20:26:58] <Bloodust> most production code is "good enough"[20:27:14] <ManDay> that's why I don't produce code. I *craft* code![20:27:16] <ManDay> :-D[20:27:25] <ManDay> also, I don't work for anyone but myself :-p[20:27:37] <jezek2> Bloodust: with list of bugs, I mean features :P[20:27:54] *** metredigm has joined ##OpenGL[20:29:31] *** Alina-malina has joined ##OpenGL[20:30:17] *** letoram has quit IRC[20:30:24] <Codex_> if I need to modify existing code, it's always some kind of failure. (maybe I found a bug that requires attention)[20:31:28] <jezek2> what about refactoring for adding some new feature? :)[20:31:29] *** samrat has quit IRC[20:31:30] *** xish has joined ##OpenGL[20:31:36] *** letoram has joined ##OpenGL[20:31:46] <ManDay> Codex_ doesn't add new features. His code has already all possible features.[20:31:47] <Codex_> new feature = independent of the existing features.[20:32:28] <jezek2> not always, sometimes you need to extend the other stuff for it's support[20:32:32] <Codex_> so once it's independent, it's just adding new code; not modifying existing code.[20:32:52] <ManDay> jezek2: Codex_' code has already all possible support for all features[20:33:08] *** MLM has quit IRC[20:33:26] <jezek2> :D[20:33:38] <Codex_> adding new code is considerably better operation than modifying.[20:33:59] <derhass> nope. the best days are whose there you delete significant amount of old code[20:34:01] <ManDay> Codex_ is our Bruce Schneider[20:34:28] <Codex_> I hope Bruce is some kind of troll[20:34:41] <ManDay> http://www.schneierfacts.com/[20:34:56] <Bloodust> bahaha[20:35:11] <Bloodust> adding more code is better than modifying old one ? :D[20:35:53] <jezek2> but I guess there is some truth in researching everything beforehand, but then it's not practical often or not possible (eg. when you don't exactly know where it would lead you for next steps, or customer changing requirements)[20:35:58] <ManDay> Bloodust: That's why you'll find a "prinf( "Hello World!" ) in all of Codex_' programs :-D[20:36:15] <jezek2> but I guess Codex_ does a bunch prototypes first when it's unknown what it's ahead :)[20:36:40] *** Zerflag has quit IRC[20:36:48] <Stragus> Real programmers never delete a line of code, they just keep on writing. *cough*[20:37:14] <Codex_> I keep dead code in the codebase obviously.[20:37:17] *** Zerflag has joined ##OpenGL[20:37:20] <jezek2> though somehow I'm not much fan of prototypes.. if it works then why not use it directly? :D[20:38:06] *** Baldwin has joined ##OpenGL[20:39:01] *** HuntsMan has quit IRC[20:39:41] *** samrat has joined ##OpenGL[20:39:45] <Codex_> i write test programs. its almost like a prototype.[20:40:56] <Codex_> its nice if i have a test case for every public function of the library. (cant fully keep that)[20:41:25] *** hexagoxel has joined ##OpenGL[20:41:29] <ManDay> Codex_: do you have a repo?[20:41:33] <ManDay> a public one, i mean[20:41:36] <Codex_> sure[20:41:42] *** Zerflag has quit IRC[20:41:59] <Codex_> http://sourceforge.net/p/gameapi/code/ci/master/tree/[20:42:42] <Codex_> GameApi.hh is the main header file[20:44:10] <ManDay> using std::placeholders::_4; … Got to love C++[20:44:29] <Codex_> yeah, that kinda sucks[20:44:47] <Bloodust> what the[20:45:14] <ManDay> your structnames are really... creative :-D[20:45:47] <ManDay> well. it's a lot of files[20:45:58] <ManDay> that's all i can say :>[20:46:33] <Bloodust> gotta conserve keyboard usage[20:46:46] *** japro has joined ##OpenGL[20:47:04] <Codex_> if its used 2 thousand times, it needs to be short.[20:47:22] <Bloodust> and obviously code completion is out of the question[20:48:05] <derhass> Bloodust: code completion? that's blasphemy![20:48:21] <ManDay> Codex_: there are 52 alpha characters. you could have given them all single letter names! imagine how short THAT would have been! :D[20:48:47] <derhass> coding is only real when you manually punch it into punchcards[20:48:48] * japro is afraid to look at the logs[20:49:06] <Bloodust> no need, this is enough http://sourceforge.net/p/gameapi/code/ci/master/tree/GameApi/GameApi.hh[20:49:46] <metredigm> those structure names are.. euh[20:49:49] <Bloodust> I should watch Highlander[20:50:09] <japro> why highlander?[20:50:14] <Bloodust> I havent seen it[20:50:18] <ManDay> lol[20:50:23] <ManDay> I expected a brilliant pun[20:50:52] <Bloodust> sorry, I couldnt come up with one :D[20:51:00] <Codex_> highlander was pretty good in 1980's[20:51:01] <japro> the one where Sean Connery is a immortal spaniard?[20:51:20] <ManDay> yeah, i remember it[20:51:20] <Bloodust> I dont know[20:51:23] <derhass> japro: there can only be one[20:51:25] <Bloodust> all I know is the quote[20:56:13] <japro> http://youtu.be/cMRgv11fIjA?t=13s[20:56:31] <japro> Juan Sanchez Villa-Lobosh Ramiresh[20:57:28] <ManDay> Oh[20:57:41] <ManDay> I think I just confused Highlander with Braveheart[20:58:05] *** Waynes1 has quit IRC[20:58:12] <japro> so, are there any games that actually use the dynamic resolution scaling from that intel aper?[20:58:13] <japro> paper[20:58:26] <ManDay> japro: If you link to it...[20:58:51] *** Orion] has joined ##OpenGL[20:59:25] <japro> the idea is fairly simple. you adjust the viewport to keep framerate stable and just scale the result to the screen[20:59:52] *** Gamecubic has quit IRC[21:00:15] <Stragus> Considering there's a cost associated with rescaling, there could be better ways to preserve performance[21:00:18] <Stragus> Levels of details, etc.[21:00:55] <japro> well i couldsee it working well in a costly deferred or otherwise fillrate bound scenario[21:01:14] <ManDay> Stragus: Rescaling is easily done and on top of that: smooth[21:01:24] <ManDay> If you change LODs you may want to have a transition etc[21:01:34] <japro> i have seen a few titles that just use different internal resolutions for their fbos[21:01:35] <ManDay> that makes things more complicated and may cost similarily many ressources[21:01:48] <japro> planetside 2 has this "quality" slider that is really just a scaler to the internal fbos[21:02:11] <Bloodust> I belive dota2 does the same thing[21:02:32] <Bloodust> it has some kind of render quality setting which makes everything look shit if you turn it down :D[21:03:14] <japro> hrhr. i think in planetside you can set the "quality" to 200% in the config file and get supersampling that way[21:03:53] <foobaz> planetary annihilation works the same way, the quality slider goes from 50% to 175% native size[21:04:40] <japro> i guess i'll try i wonder how noticeable the resolutions shifting is[21:04:57] *** metredigm has quit IRC[21:06:03] <Bloodust> should try it[21:06:59] *** seangrove has quit IRC[21:07:33] *** seangrove has joined ##OpenGL[21:14:46] *** biberao has joined ##OpenGL[21:14:48] <biberao> hi[21:15:11] *** Gamecubic has joined ##OpenGL[21:17:57] <Bloodust> hi[21:18:10] *** Vtec234 has left ##OpenGL[21:19:12] <biberao> anyone uses python for opengl?[21:20:03] <Codex_> no idea[21:20:27] *** samrat has quit IRC[21:20:37] <foobaz> not me[21:21:10] <Bloodust> nope[21:21:30] <japro> i guess there needs to be some clever filtering[21:21:47] <japro> i just implemented it via chaing Viewport and framebuffer blit[21:21:56] <japro> looks terrible[21:22:34] <ManDay> japro: lol, ofc[21:22:50] <ManDay> did you just upscale the viewport with a linear filter or what?[21:23:12] <Bloodust> he[21:23:45] <derhass> use nearest sampling for the retro effect[21:24:28] <derhass> retro is in![21:26:03] <japro> yep, :D nearest 4x upscale[21:26:13] <japro> that might almost qualify as "art style" then[21:26:51] <japro> there is no way to use the default framebuffers stencil buffer with a fbo right?[21:27:24] <japro> i need the stencil in my defered resolve but that means i have a pointless 1:1 blit to the screen that costs me almost a milisecond[21:27:48] <biberao> anyone here doest opengl for the rpi?[21:28:00] <japro> i have done that[21:28:42] <biberao> japro: you're answering me[21:28:44] <biberao> ?[21:28:52] <japro> yep[21:29:26] <japro> beat it about half a year ago... the end boss was hard[21:29:43] <biberao> im trying to understand about it[21:29:50] <biberao> i wanted to do this[21:30:11] <biberao> https://www.youtube.com/watch?v=4iwQAqgG1uI[21:34:44] *** ManDay has quit IRC[21:35:55] <japro> so, what is the question? :)[21:36:50] <japro> amusingly the rpi can actually put a transparent gl surface on top of it's console without x running :D[21:36:56] <japro> so you could run a shader on top of vi or so? :D[21:37:01] <biberao> really?[21:37:06] <biberao> thats possible>?[21:37:22] *** Zerflag has joined ##OpenGL[21:37:26] <biberao> and have it running realtime like that and from ssh to lets say another device[21:37:29] <biberao> like a tv?[21:37:45] <foobaz> you can do that on regular PC linux too[21:37:56] <biberao> i believe so[21:38:04] <biberao> but wanted to do some nice stuff for a tv[21:38:26] <japro> foobaz, don't you need something funky like weyland for that though?[21:38:48] <foobaz> you just need EGL and KMS[21:38:58] <foobaz> that is how wayland works[21:39:08] <foobaz> but you can use it to get a raw gl context without using wayland[21:39:16] <derhass> that's not the same[21:39:25] *** zzuegg has quit IRC[21:39:42] *** paul424 has joined ##OpenGL[21:39:42] <derhass> you'll still need a compositor[21:39:49] <derhass> on the pi, you don't[21:40:02] *** ablephar` has joined ##OpenGL[21:40:31] *** zzuegg has joined ##OpenGL[21:41:23] <japro> so all you need is a gl program that recompiles and runs shaders on uh inotify or so and need to figure out the firmware api of the rpi to get that transparent window[21:41:38] *** shintahW2 has joined ##OpenGL[21:42:18] *** ablepharus has quit IRC[21:42:51] *** Biliogadafr has quit IRC[21:43:26] <derhass> japro: you don't need gl at all[21:44:43] <derhass> but yeah, reloading shaders via inotify could be funny[21:44:49] *** shintahW has quit IRC[21:44:50] *** shintahW2 is now known as shintahW[21:45:13] <japro> derhass, ?[21:45:37] <japro> if the purposes is to livecode glsl shaders you can avoid gl?[21:45:45] <derhass> nope[21:47:27] <japro> interesting running fxaa on a pixel checkerboard is really friggin expensive :D[21:47:48] *** samrat has joined ##OpenGL[21:47:51] <derhass> hrhr[21:48:34] <japro> 10ms vs ~2.5 on a normal scene[21:51:12] *** Zerflag has quit IRC[21:52:55] <biberao> im back sorry[21:52:55] *** ManDay has joined ##OpenGL[21:53:27] <foobaz> japro: what is the result? uniform grey?[21:53:36] <japro> foobaz, close to it[21:53:54] <foobaz> i never liked FXAA, just seems to make things blurry[21:54:31] <japro> i find it acceptable for very vectory stuff (which i tend to do not being an artist...)[21:54:44] <foobaz> and it doesn’t actually fix aliasing like with far away power lines, fences, or other thin objects[21:54:57] <biberao> https://benosteen.wordpress.com/2012/04/27/using-opengl-es-2-0-on-the-raspberry-pi-without-x-windows/[21:55:01] <japro> also i googled this earlier and found some really blurry screenshots where i thought "my fxaa implementation doesn't look like that???"[21:57:03] <biberao> i wanted to have ipython on the fb or somehow hehe[21:57:05] *** HuntsMan has joined ##OpenGL[21:57:36] <biberao> or actually have the livecoding part from ssh and output to the other device in realtime[21:58:25] <japro> foobaz, http://i.imgur.com/EYwkDLM.png?1 looks ok to my eyes[21:58:48] <biberao> any idea?[21:59:30] <foobaz> japro: ah yea that is very simple geometry, i can see how it would work better on that[21:59:34] *** karab44 has quit IRC[21:59:40] <foobaz> your game looks a lot like “race the sun”, ever play that?[22:00:01] <foobaz> http://store.steampowered.com/app/253030/[22:00:10] <japro> uh, yeah :D[22:00:26] <japro> although i kinda got it because i started working on this[22:00:47] *** foreignFunction has quit IRC[22:02:04] *** realz has quit IRC[22:02:36] *** MLM has joined ##OpenGL[22:02:39] *** realz has joined ##OpenGL[22:03:27] *** snakenerd has quit IRC[22:04:34] <Bloodust> japro sweet[22:06:01] <danhedron> nice plane[22:06:08] <danhedron> 11/10 would avoid geometry[22:09:24] <Codex_> japro: how did you create the plane model?[22:09:59] *** jdolan has quit IRC[22:11:05] <japro> the fuselages are made in blender and the rest is done "programmatically". so i have a bunch of functions that create parametrized shapes such as wings, rotation profiles etc.[22:11:54] *** Keniyal has joined ##OpenGL[22:13:20] <japro> foobaz, i couldn't figure out how race the sun does shadows.... i guess unity just uses shadowmaps. but it seems it only shadowmaps the glider[22:13:40] <japro> enviroment shadows seem to vary between nonexistent and glitched[22:14:25] <japro> maybe because i was playing on a toaster[22:14:55] <derhass> a flying toaster?[22:15:21] <japro> a netbook ;)[22:15:39] * foobaz tags derhass as “old”[22:15:56] <derhass> foobaz: I _am_ old[22:16:13] <derhass> I can't keep up with the cool kids here[22:16:14] <foobaz> i used to have an inflatable flying toaster, swag from after dark :)[22:17:38] <dahlia> with flapping wings?[22:17:50] <foobaz> maybe if you flapped them yourself[22:18:04] * dahlia flaps...[22:18:50] *** SugarCube has quit IRC[22:19:43] *** SugarCube has joined ##OpenGL[22:20:35] *** disappointment has joined ##OpenGL[22:22:18] *** mizux has joined ##OpenGL[22:23:42] <Codex_> netbook sounds like very bad idea for opengl development.[22:24:26] <foobaz> on the plus side, if you can make it run on a netbook, it will run anywhere with great performance[22:24:39] <japro> true :) although i was positively surprised by it since its at least a last generation intel one[22:25:25] <Codex_> i had eeepc and it didnt work well for development with intel card[22:25:38] <foobaz> gma 3150?[22:25:39] *** disappointment has quit IRC[22:25:46] <japro> yeah i had an older netbook too that was horrible[22:25:49] <foobaz> that gpu was crap[22:26:11] *** TyrfingMjolnir has quit IRC[22:26:12] <japro> this one has the same intel HD as the uh, ivy bridge cpus i think?[22:26:51] <japro> celeron N2830 the thing from the screenshot earlier runs just about 60hz at 800x600[22:27:09] <japro> take that consoles!!![22:27:11] <japro> :D[22:27:32] <derhass> I still have an old netbook[22:27:36] *** stefkos has quit IRC[22:27:51] <foobaz> my laptop is 11”, but it has a core i7 :)[22:28:21] <Codex_> the eeepc also needed awful hacks to get linux installed[22:29:17] <Codex_> since the manufacturer nicely filled their partition table, and you needed to erase one hard disk to get linux in... default install didnt work[22:30:24] *** tehrain has quit IRC[22:30:38] *** konom has quit IRC[22:31:01] <Codex_> (otoh, thats better than my current laptop, where linux installation was impossible)[22:38:09] *** samrat has quit IRC[22:39:54] *** razieliyo has joined ##OpenGL[22:39:59] *** razieliyo has joined ##OpenGL[22:41:38] *** meoblast001 has joined ##OpenGL[22:42:38] *** disappointment has joined ##OpenGL[22:46:39] *** BreadProduct has joined ##OpenGL[22:49:42] *** Vtec234 has joined ##OpenGL[22:54:53] *** japro has quit IRC[22:57:14] *** BearishMushroom has quit IRC[23:01:19] *** Zerflag_ has joined ##OpenGL[23:04:43] *** Lemml has quit IRC[23:06:57] *** anivemin has quit IRC[23:08:32] *** roxlu has quit IRC[23:08:38] *** zagabar has quit IRC[23:11:11] *** zagabar has joined ##OpenGL[23:11:11] *** zagabar has joined ##OpenGL[23:13:15] *** garFF has joined ##OpenGL[23:16:16] *** BitPuffin has joined ##OpenGL[23:18:15] <ManDay> Codex_: dd if=/dev/zero of=/dev/sda count=10000[23:18:18] <ManDay> there you go[23:19:15] *** jdolan has joined ##OpenGL[23:19:26] *** slidercrank has quit IRC[23:19:28] <Codex_> no go[23:19:43] <Codex_> it still wouldnt boot to linux.[23:20:14] <foobaz> did you install a boot menu item with efibootmgr?[23:23:06] <Codex_> never got to that far[23:23:52] *** razieliyo has quit IRC[23:24:05] *** t4nk093 has joined ##OpenGL[23:24:32] *** telex has quit IRC[23:24:33] *** Demon_Fox has joined ##OpenGL[23:25:05] *** konom has joined ##OpenGL[23:26:24] *** razieliyo has joined ##OpenGL[23:26:24] *** razieliyo has joined ##OpenGL[23:26:38] *** telex has joined ##OpenGL[23:27:01] *** ManDay has quit IRC[23:27:21] *** Vtec234 has quit IRC[23:31:44] *** Miklas has quit IRC[23:32:32] <roboman2444> Codex_, sounds like it was an efi device[23:32:42] <roboman2444> bios settings should fix it[23:32:58] *** magicthing has quit IRC[23:33:02] <t4nk093> is there a good tutorial on texture atlases?[23:33:09] <t4nk093> for java[23:33:14] *** phaazon has quit IRC[23:33:23] *** elveni has quit IRC[23:33:28] *** petris has quit IRC[23:34:07] <Codex_> tank093: you just put all your sprites to same bitmap, and then use texture coords to select correct sprite?[23:34:43] <Codex_> I havent seen good tutorial on them[23:36:07] <t4nk093> the problem is i need to use array textures with the atlas[23:36:27] <roboman2444> do the same thing, but now pass a third TC as a int[23:36:29] <roboman2444> for the z coord[23:36:30] <t4nk093> the atlas is just for UI and stuff with images that aren't all the same size[23:36:42] *** petris has joined ##OpenGL[23:37:48] <t4nk093> how do you have the shader distinguish between the 2 types?[23:39:02] *** kuldeepdhaka has quit IRC[23:39:18] *** qeed has joined ##OpenGL[23:41:13] *** Miklas has joined ##OpenGL[23:42:19] <t4nk093> anyone have any idea how to use both at the same time?[23:44:54] <derhass> at the same time?[23:45:30] *** Misu has quit IRC[23:46:06] *** bb010g has joined ##OpenGL[23:47:45] *** linuxuz3r_ has joined ##OpenGL[23:50:18] *** phao has joined ##OpenGL[23:50:31] <phao> Hi. I don't mean to sound too pessimistic or to start any flame wars or anything...[23:50:50] *** linuxuz3r has quit IRC[23:50:50] <phao> but is it really that the best materials to learn opengl are the red book and that arcsynthesis guide?[23:50:57] <phao> Nothing better out there?[23:51:57] <slime> what's best for you depends on how you learn and what you already know[23:52:40] <phao> Hi slime[23:52:49] <slime> hi[23:53:17] <roboman2444> open.gl[23:54:42] <phao> slime, I don't know much of opengl.[23:55:32] <phao> I've done some fixed function programming, and that's it.[23:55:38] <phao> I've came to use a shader once though...[23:57:28] <slime> there's also the GL superbible, and various other resources online and in print[23:58:00] <t4nk093> do i need to render array texture objects and texture atlas objects with seperate draw calls?[23:58:01] <slime> personally the best way I learn is with a combination of documentation, books or tutorials, and existing open source code I can look at and tinker with[23:58:22] <t4nk093> after binding the correct texture?[23:58:23] *** navitsu has quit IRC[23:58:37] <phao> slime, ok[23:58:44] <t4nk093> or is there a way to avoid that?[23:58:47] <phao> for a beginner, the superbible is more friendly than the redbook?[23:59:01] <slime> phao: probably[23:59:09] <phao> I didn't know that. Thanks.[23:59:45] <Codex_> tank093: I thought your shader would just choose correct texture from the array.